This paper deals with the training procedure for a hierarchical neural network (Tree of Multi-Layer Perceptrons—TMLP) aimed to classify surface defects in flat rolled strips. Due to the difficulties in collecting large Data Bases it is necessary to exploit at the best the available knowledge. A comparison between techniques derived from both the Back-Propagation and Weight-Perturbation algorithms is done, and experimental results are reported.
Gradient descent learning algorithm for hierarchical neural networks: A case study in industrial quality
VALLE, MAURIZIO;CAVIGLIA, DANIELE
1999-01-01
Abstract
This paper deals with the training procedure for a hierarchical neural network (Tree of Multi-Layer Perceptrons—TMLP) aimed to classify surface defects in flat rolled strips. Due to the difficulties in collecting large Data Bases it is necessary to exploit at the best the available knowledge. A comparison between techniques derived from both the Back-Propagation and Weight-Perturbation algorithms is done, and experimental results are reported.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
